应用第二代Curvelet变换的遥感图像融合

Fusion of remote sensing images based on the second generation Curvelet transform

提出了一种基于第二代Curvetlet变换遥感图像融合算法.将具有高空间分辨力的Pan图像与Ms图像的待融合波段图像进行直方图匹配,并对直方图匹配后的Pan图像与待融合波段Ms图像分别进行Curvelet变换分解,得到各自的低频子带系数和各带通方向子带系数;采用一定的融合规则对Curvelet变换系数进行组合得到融合图像的Curvelet系数;最后对组合后的系数进行Curvelet重构得到该波段具有高空间分辨力的Ms图像.对IKONOS卫星遥感图像的仿真实验结果表明:与传统的基于亮度-色调-饱和度彩色空间变换融合算法相比,该算法使融合后的Ms图像整体光谱保持度提高了10.54%,而与传统的基于小波变换的图像融算法相比,其空间质量提高了0.81%～1.12%,有效解决了基于亮度-色调-饱和度彩色空间变换融合算法中光谱失真严重和基于小波变换图像融合算法中空间质量较低的缺点,使得融合后的Ms图像在最大可能地保持原始Ms图像光谱特性的同时,显著提高了融合图像的空间质量.

A novel fusion algorithm for remote sensing image based on the second generation Curvelet transform is proposed. The histogram matching between the Pan image with high spatial resolution and the band image to be fused in the Ms images is performed to eliminate the spectral differences, and the histogram matched Pan image and the band image to be fused in the Ms images are decomposed by the Curvelet transform. Then, some fusion rules are employed to obtain the Curvelet coefficients of the fused image. Finally, the inverse Curvelet transform is performed to obtain the fused Ms image with high spatial resolution. Using the proposed method to simulate several sets of IKONOS images, experimental results demonstrate that the total spectrum preservation is improved by 10.54% compared with the Intensity-hue-saturation (IHS) transform method and the spatial quality is improved by 0.81%-1.12% compared with the Discrete Wavelet Transform (DWT) method. The proposed algorithm also can overcome the high spectrum distortion of the IHS method and the low spatial quality of the DWT method, so that the proposed algorithm can not only improve the spatial quality of the fused Ms images effectively, but also make the fused images in little spectrum distortion.